Convulsive status epilepticus (CSE), with an incidence of 17–23 episodes per 100 000 children per year,1 is the most common medical neurological emergency in children. Since there is significant associated morbidity and mortality, which in part may be related to seizure length,2 it is essential that acute paediatric and paediatric emergency staff are comfortable and familiar with its management.

To maximise the probability of seizure termination, protocols for early appropriate treatment need to be developed. This review aims to examine the justification for early treatment and evidence supporting certain therapeutic interventions and to identify similarities and differences in protocols worldwide to identify and promote best practice.

DEFINITION

CSE is often defined as either two or more convulsions without complete recovery of consciousness between seizures (intermittent CSE) or a single prolonged seizure lasting at least 30 min (continuous CSE) with a motor component. The motor component usually consists of tonic stiffening followed by clonic movements of all limbs, although purely tonic or clonic as well as myoclonic status epilepticus can occur.3 There has been a move towards an operational definition of CSE for treatment purposes as any seizure lasting longer than 5 min, which will be discussed later.4

DIAGNOSIS

The diagnosis of overt tonic-clonic seizures is usually straightforward. The main differential diagnoses are non-epileptic events (previously known as pseudoseizures) and other causes of abnormal movements such as acute dystonia, paroxysmal dyskinesia or involuntary movements associated with cerebral palsy. While uncommon, children with non-epileptic events can appear to be in CSE, but such children usually have a background of coexisting epilepsy (which may or may not have been recognised) or a family history of epilepsy.5 ,6 Therefore, in the acute situation it will usually be preferable to treat for CSE unless background information is available, although it is equally important to have a high threshold for using CSE protocols to treat individuals with known recurrent non-epileptic events.

In the absence of emergency electroencephalography, other laboratory investigations are rarely helpful in diagnosing true epileptic seizures but may be helpful in identifying aetiology. There is conflicting evidence as to the value of blood tests, such as serum prolactin, serum creatine kinase and white blood count levels, in differentiating between epileptic and non-epileptic seizures, but it is recommended that these measurements are not used for the diagnosis of epileptic seizures.7 An acute rise in plasma lactate can be seen during seizures8 but can also be seen as a result of strenuous physical activity, limiting its diagnostic role.

In children with pre-existing neurodisabilities, especially if they are prone to abnormal movements, it is difficult to differentiate more subtle seizure types. Treatment decisions can be difficult if staff are unfamiliar with the child, but carers’ views and pre-agreed treatment plans or individualised protocols can help guide management.

AETIOLOGY

Although the underlying cause of a seizure does not affect the primary assessment and immediate management of CSE, it is the main determining factor for further treatment, investigation and eventual prognosis.2 ,9 A distinction between febrile and afebrile CSE is useful. A child’s temperature can be measured on arrival in hospital, but since body temperature can rise during a seizure as a result of muscle activity, acute temperature values in a child with a prolonged seizure should be interpreted with caution. A recorded temperature before seizure onset is ideal, but this is not always available. A child is considered febrile if they have a body temperature of over 39.0°C (38.0°C if under 3 months of age) measured by either an electronic or chemical dot thermometer in the axilla or an infra-red tympanic thermometer.10 However, a temperature of over 38.0°C is considered appropriate for a definition of febrile CSE.

The main differential diagnoses of febrile CSE are central nervous system (CNS) infections, co-incidental febrile illness in a child with a pre-existing condition that predisposes to CSE (eg, a child with cerebral palsy with pre-existing epilepsy) and prolonged febrile seizures (PFS).11 Children with CNS infections mismanaged as PFS are at risk of death and severe neurological sequelae and therefore PFS should be a diagnosis of exclusion.11 See box 1 for the main differential diagnoses for afebrile CSE.

Children with afebrile CSE can be considered in two groups: those previously neurologically normal and those with pre-existing neurological concerns/conditions including those with developmental delay. CSE may be the first seizure in a child who is subsequently diagnosed with epilepsy, be the second seizure that leads to the diagnosis of epilepsy or may occur in a child with pre-existing epilepsy. Investigations should be targeted at narrowing the range of differential diagnoses and should be guided by clinical history and examination.

Low anti-epileptic drug (AED) levels are a common cause of CSE in adults but an uncommon cause in children.1 Therefore, care should be taken to exclude other causes of CSE in children with known epilepsy. Other than drugs of abuse, particular substances associated with toxic ingestion or overdose leading to CSE include theophylline, lindane, carbamazepine and chemotherapeutic agents.12 These need specific tests if suspected.

PROTOCOLS AND GUIDELINES

The aims of protocols are to ensure appropriate and effective treatment of common conditions or situations. In the stressful environment of a medical emergency, familiarity with the recommended management saves time and prevents errors. Many different treatment guidelines and protocols have been proposed in an attempt to improve the management of CSE. Due to the limited evidence base, the majority are based on a combination of local experience, consensus opinion and local drug availability and it is hard to recommend any particular protocol as a “gold standard”. Since the words “clinical guideline” currently imply a certain status and methodology,13 we have used the term “protocol” to refer to suggested treatment algorithms except where the source meets clinical guideline criteria.

Despite the proliferation of protocols for CSE, studies and local audits repeatedly find they are often not followed, resulting in longer CSE and higher rates of admission to intensive care.14^–18 The main problem areas are:

• 1) giving inadequate doses of benzodiazepines, particularly rectal diazepam,15 ,16

• 2) treating with more than two doses of benzodiazepines,15 ,18

• 3) delay in initiating second line treatment and/or anaesthetic intervention.14

In some individuals there may be good reasons for deviating from established protocols. Children with recurrent CSE unresponsive to AEDs administered according to established protocols may benefit from individualised protocols drawn up with their responsible clinician and distributed to local services and parents.

The main UK protocols are that of the Advanced Life Support Group in Advanced paediatric life support (APLS)9 and that suggested by the National Institute for Health and Clinical Excellence (NICE) in 2004 as part of their clinical guidelines for epilepsy (an updated publication is also available).19 Both of these are heavily based on the work of the Status Epilepticus Working Party (SEWP) in 2000.20

Other publicly available protocols include those published for use in New South Wales,21 Boston Children’s Hospital,22 Starship Children’s Hospital in New Zealand23 and Japan.24

A generic treatment protocol is given in fig 1. Initial hospital treatment usually depends on whether intravenous access can be obtained quickly, and then proceeds down the flowchart until seizure termination, typically with recommended pauses of 10 min between treatments. This may inevitably cause a delay of up to 40 min before anaesthetic intervention is initiated. Most protocols take this form, with the main differences being in recommended drugs, drug dosage and treatment order (see ⇓tables 1 and 2). Stated drug doses are for individual protocols and the British National Formulary for children recommended dosing for listed AEDs is also given for reference. The evidence behind these recommendations is reviewed below.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Generic treatment protocol for convulsive status epilepticus. Grey shaded boxes do not feature in all guidelines. IV, intravenous; PR, per rectum.

PRINCIPLES OF TREATMENT

The aim of treating CSE is to prevent brain injury from ongoing seizures. Although the underlying aetiology is probably the primary determinant of short and long term outcome, prolonged seizure activity per se is widely believed to be harmful.25

In animal models, ischaemic and excitotoxic cell loss26 starts to occur after 30 min of seizure activity27^–29 and worsens with increasing seizure duration. Case reports of CSE show progressive signs of cerebral atrophy on magnetic resonance imaging (MRI), associated with post-mortem evidence of neuronal loss.30^–32

Although most seizures self-terminate within 5 min, seizures that last longer than 5–10 min are at high risk of continuing for at least 30 min.33 ,34 As a seizure continues it also becomes progressively more difficult to stop,15 ,35 and so the best chance of terminating a seizure is with early treatment. This is the reasoning behind the operational definition of CSE described earlier, with emergency treatment starting at 5 min.

Unless a seizure starts in hospital, this will probably involve initiation of treatment by parents or first responders. This should start with a rapid assessment and optimisation of the patient’s airway, breathing and circulation, remembering that the best way of improving these will often be with early AED administration. Hypoglycaemia is an important and easily correctable cause of CSE and early blood glucose measurement is important.

FIRST LINE TREATMENT

Benzodiazepines are the usual first line drugs for treating CSE. They have good efficacy and rapid action, and are easy to prepare and administer. Intravenous, rectal or buccal preparations of commonly used benzodiazepines (ie, diazepam, lorazepam and midazolam) are readily available in most countries.

Benzodiazepines primarily work by binding at the GABA_A receptor and promoting neuronal inhibition.36 ,37 Therefore as an episode of CSE progresses, the number of active GABA_A receptors decreases38 and they become less effective.39 ,40 If the first dose of benzodiazepines is given early (within 20 min of seizure onset) trials have shown 70–86% seizure termination, but this decreases significantly as the seizure progresses.41

In the pre-hospital setting, intravenous access is unlikely to be available. Rectal diazepam has been the traditional treatment of choice,41 however recent studies have shown that buccal midazolam is a safe and more effective choice across a range of settings. It is also easier to administer, more socially acceptable and preferred by parents.42^–45

Once in hospital, intravenous access should be obtained as soon as possible, as it offers superior onset of action, bioavailability and efficacy.15 ,46 When available, intravenous lorazepam is usually the drug of choice. It has longer lasting anticonvulsant activity than diazepam with a similar risk of respiratory depression. A recent Cochrane review concluded that intravenous lorazepam was safer and at least as effective as intravenous diazepam47 and a recent population based study found that after adjusting for seizure, patient and other treatment characteristics, intravenous lorazepam was almost four times more likely to be associated with seizure termination compared to rectal diazepam as initial treatment on arrival at the accident and emergency department.15

Should the first AED prove ineffective, then further treatment should be started. Lorazepam takes 5–10 min to be effective (median time 6 min) and most protocols incorporate a 10 min observation period between doses of AEDs.46 If the seizure continues after this period, most protocols suggest giving at least one further trial of benzodiazepines, although the evidence suggests that this is less likely to prove effective as there is much lower response rate to the second dose of diazepam (16.7% compared to 85% for the first dose in one study).48

Seizures that have not responded to two doses of benzodiazepines are much less likely to respond to further doses15 as this will inevitably be later into the seizure. There is also an increased risk of side effects with multiple doses, especially respiratory insufficiency.14 ,49 There should be a rapid move to second line treatments once initial treatment with benzodiazepines has failed. Therefore, if children have received benzodiazepines in the pre-hospital setting, then perhaps only one dose of intravenous benzodiazepine should be administered in hospital, contrary to current advice.

There is little reported evidence on the use of acute doses of benzodiazepines in children already on maintenance benzodiazepines. Animal models suggest that tolerance may be drug specific in at least some cases,50 so there may be a response to acute administration of a single dose of benzodiazepine, but a low threshold should exist for moving onto other treatments should this fail.

SECOND LINE TREATMENT

The most commonly proposed drugs for second line treatment are paraldehyde, phenytoin and phenobarbital. There have been no randomised controlled trials of second line therapies47 so choices are based on other evidence.20

REFRACTORY SEIZURES

CSE unresponsive to two different anti-epileptic medications is considered refractory (some authorities add a >60 min duration criterion) and unlikely to respond to further doses of usual AEDs.64 ,72 Treatment of refractory CSE should be with senior advice and anaesthetic support and is beyond the scope of this review, except to note that delays should be minimised.

CONCLUSION

With limited available data on the treatment of CSE, it is difficult to offer more than a few basic principles for the gold standard treatment of this condition. Its time critical nature is probably the most important factor73 rather than the exact form of individual interventions. For community onset seizures, pre-hospital treatment should be maximised and taken into account by hospital personnel. Acute paediatric staff should be familiar with the local protocol for the treatment of CSE so that appropriate rapid treatments are given.

While all the medications listed above have been shown to have some anti-seizure activity in CSE, an optimal line of treatment is emerging as follows.

• Benzodiazepines, of which buccal midazolam and intravenous lorazepam are superior, are the best first line treatment.

• If two doses of benzodiazepines (pre-hospital doses should be regarded) have not been effective, then rapidly switch to second line agents. Intravenous phenytoin is far more effective than rectal paraldehyde. Other treatment options are available, but further research is needed.

• If two different anti-epileptic agents fail to stop seizure activity, referral for intubation and anaesthesia with definitive treatment for refractory status epilepticus should be considered after 20 min.

We propose a protocol (fig 2) that takes the above points into account. This is similar to one we previously published,74 with the inclusion of a single dose of buccal midazolam if intravenous access is not achieved quickly after arrival in hospital. While this and other currently available protocols are helpful, they should be regarded as being in a constant state of revision. As new evidence emerges, protocols and guidelines will need revising and updating. Two of the authors (RFC and RCS) are members of the SEWP which is currently reviewing its protocol.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Proposed treatment protocol for convulsive status epilepticus. IO, intraosseous; IV, intravenous; PE, phenytoin equivalent.

Protocols provide a baseline for treatment on a population basis, in the absence of specific knowledge. Particular local expertise or limitations may also provide legitimate reasons to adjust or adapt protocols. For individual children who are known to respond well to specific medications, a more tailored approach is recommended.